Summary
This Reuters Institute report provides one of the clearest empirical baselines in the archive for how working journalists are actually using AI. Based on a broadly representative survey of 1,004 UK journalists, it shows that AI use had already become routine for many by late 2024, especially in transcription, translation, grammar checking, story research, and some verification work.
Why It Matters
This is a strong legacy direct story because it replaces general newsroom chatter with task-level adoption data.
- it distinguishes between low-risk support uses such as transcription and more substantive tasks such as research, headline generation, and source assessment
- it shows that AI use is uneven across beats and responsibilities, which matters for newsroom policy design and training
- it documents that journalists were using AI much more for language and information-processing tasks than for image or video generation
What the Source Says
The report says 56% of UK journalists use AI professionally at least once a week, another 27% use it less frequently, and 16% have never used it for journalistic tasks. It reports especially high monthly use for transcription or captioning, translation, and grammar checking or copy-editing. It also finds meaningful use in story research, brainstorming, headline generation, and fact-checking or source assessment, while image and video generation remain rare. The authors further note that journalists with management responsibilities and journalists on some beats, such as business coverage, use AI more often than others.